This study examines how warnings accompanying televised beer advertisements affect viewer perceptions of the risks and benefits of alcohol consumption, as well as how much warnings influence viewers' responses to the beer advertisements themselves. This research will also evaluate alternative warning content and wording, placement, and audio-visual presentation. Our goal is to provide a validated theoretical model, from the perspective of persuasion, attitude, and behavior change, describing how warnings may influence alcohol risk and benefit judgments and other behavioral antecedents, and to provide an empirical framework for recommendations regarding the content and presentation of such warnings. Phase 1 will provide a test of various warning content strategies, including use of base-rate (statistical) information, and provision of behavioral recommendations; alternative warning approaches will be pretested in focus groups. Criteria will include subjective cognitive responses, and impact on knowledge, risk estimates, self-efficacy, and intention. Phase 2 will test preferred warning strategies against a no-warning control condition, and manipulate presentation of warnings before or after ad presentation; dependent variables will also include perceptions of alcohol risks and benefits and confidence in those perceptions. Phase 3 will examine audio- visual content issues, including manipulating text-only, positive-appeal visuals (e.g., beer being poured), and fear-inducing visuals (e.g., automobile accident footage), crossed with no-audio, authoritarian, and non-threatening vocal quality voice overs. Random digit dialing recruiting will be used throughout, to maximize external validity. Total N will be 800-1000 for Phases 1-3. Subjects in all three phases will be non- abstinent adults ages 21-65, stratified by gender and amount of drinking.